The Next Byte Podcast · 11 Jul 2023 · 18m
Daniel and Forbode discuss a non-invasive hydrogel patch developed by researchers at ETH Zurich and EMPA to detect leaks after abdominal surgery. The patch uses tantalum oxide to react with gastric fluids, creating CO2 bubbles detectable via CT or ultrasound, which are then analyzed by a computer vision algorithm. The hosts highlight how this technology addresses the high mortality rate of post-surgical leaks and could reduce hospital stays and healthcare costs.
Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.
Engineering & simulationPython algorithm
A Python algorithm was connected to computer vision to algorithmically evaluate CT and ultrasound images of the hydrogel patch to determine if leakage had occurred, removing human error from the assessment.
“they connected this using computer vision to a Python algorithm, um, that can algorithmically evaluate the CT scans and the ultrasound images to tell whether or not there has been leakage on the gut.”
Speaker 1 · Hear it at 7:59
makers and hobbyists
Other shows on the same ground, closest first.
Professor Dan Stoyanov of University College London and co-founder of Odin Vision discusses the convergence of surgical robotics, computer vision, and AI. He explains how AI assists in pre-operative planning, real-time surgical guidance (such as polyp detection in gastroenterology), and post-operative patient monitoring. Stoyanov details the founding of Odin Vision as a UCL spin-out to deploy these AI systems clinically and reflects on the challenges of transitioning from academic research to industrial application.
Hackaday editors Elliot Williams and Dan Maloney review a diverse range of hardware hacks, including a German researcher's software patch for encrypted medical connectors, a Starlink signal reverse-engineering project for GPS, and a mechanical forearm muscle sensor using Hall effect sensors. They also discuss a University of Maryland project creating flexible PCBs via laser ablation, a software decompilation hack to fix an obsolete solar inverter, and an explanation of Stable Diffusion AI. The episode concludes with a Commodore 64-based accordion, a citizen science water colorimeter, and a discussion on a Glasgow nightclub's heat recovery system.
Dr. Gregory Charvat, CTO of Humatics Corporation, discusses his three startups: Butterfly Network, Hyperfine, and Humatics. He details the pivot at Butterfly Network from high-cost ultrasound tomography to a handheld ultrasound on a chip using MEMS technology. He also describes Humatics' development of a short-range, high-precision indoor GPS system for factory automation, which achieved centimeter and millimeter accuracy through time-of-flight measurements.
Hackaday editors Elliot Williams and Tom Nardi discuss several hardware projects, starting with the security vulnerabilities of Apple's AirTag tracking system, where Fabian Boyne Lines demonstrates how an ESP32 can spoof the network to evade detection. They then cover DIY hydrophones built by DJ Jules using piezoelectric transducers and 3D printed 'warped' objects created by Stefan from CNC Kitchen by modifying G-code. The episode highlights Jhui Jun's high-precision DIY robot arm capable of suturing grapes, and features a deep dive into the petrochemical origins of wintergreen flavoring and the history of the Canadian MCM 70 portable computer.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Next Byte Podcast. The episode itself belongs to its publisher.